Hair pattern determination and filtering

Abstract

Described are systems and methods to determine hair patterns presented in content items. The determined hair patterns may be associated with the content items to facilitate indexing, filtering, etc. of the content items based on the determined hair patterns. In exemplary implementations, a corpus of content items may be associated with an embedding vector that includes a binary representation of the content item. The embedding vectors associated with each content item can be provided as inputs to a trained machine learning model, which can process the embedding vectors to determine one or more hair patterns presented in each content item while eliminating the need for performing image pre-processing prior to determination of the hair pattern(s) presented in the content item.

What is claimed is: 
     
       1. A computing system, comprising:
 one or more processors; and 
 a memory storing program instructions that, when executed by the one or more processors, cause the one or more processors to at least:
 access a plurality of embedding vectors, wherein each embedding vector of the plurality of embedding vectors is representative of a corresponding content item of a corpus of content items; 
 process, using a trained machine learning model, each embedding vector of the plurality of embedding vectors to determine a respective hair pattern for each corresponding content item; 
 associate each respective hair pattern with each corresponding content item; 
 determine, based at least in part on a query received from a client device, a first plurality of content items from the corpus of content items that are responsive to the query; 
 cause, in response to the query, a first user interface to be presented on the client device, wherein the first user interface includes at least a portion of the first plurality of content items and a hair pattern filter control that includes a plurality of selectable hair patterns; 
 obtain, via an interaction with the hair pattern filter control, a selection of a first hair pattern from the plurality of selectable hair patterns; 
 determine a second plurality of content items from the first plurality of content items, wherein each of the second plurality of content items is associated with the first hair pattern; and 
 cause, in response to the selection of the first hair pattern, a second user interface to be presented on the client device, wherein the second user interface includes at least a portion of the second plurality of content items. 
 
 
     
     
       2. The computing system of  claim 1 , wherein the program instructions include further instructions that, when executed by the one or more processors, further cause the one or more processors to at least:
 determine that the query triggers diversification of responsive content items. 
 
     
     
       3. The computing system of  claim 2 , wherein at least the portion of the first plurality of content items is presented on the client device according to a diversity associated with the first plurality of content items. 
     
     
       4. The computing system of  claim 2 , wherein at least the portion of the second plurality of content items is presented on the client device according to a diversity associated with the second plurality of content items. 
     
     
       5. The computing system of  claim 1 , wherein the program instructions include further instructions that, when executed by the one or more processors, further cause the one or more processors to at least:
 prior to causing the first user interface to be presented on the client device, determine that the query triggers hair pattern filtering. 
 
     
     
       6. The computing system of  claim 1 , wherein:
 the program instructions include further instructions that, when executed by the one or more processors, further cause the one or more processors to at least obtain, via a second interaction with the hair pattern filter control, a second selection of a second hair pattern from the plurality of selectable hair patterns; and 
 determining the second plurality of content items from the first plurality of content items is further based on the second hair pattern, such that each of the second plurality of content items is associated with at least one of the first hair pattern or the second hair pattern. 
 
     
     
       7. A computer-implemented method, comprising:
 processing, using a trained machine learning model, a plurality of embedding vectors associated with a corpus of content items to determine corresponding hair patterns for the content items included in the corpus of content items; 
 associating the corresponding hair patterns to the content items included in the corpus of content items; 
 determining, in response to a query received from a client device, a first plurality of content items from the corpus of content items that are responsive to the query, wherein each content item of the first plurality of content items is associated with a respective hair pattern from the corresponding hair patterns; 
 causing at least a portion of the first plurality of content items and a hair pattern filter control to be presented on the client device, wherein the hair pattern filter control presents a plurality of selectable hair patterns; 
 obtaining an interaction with the hair pattern filter control selecting a first hair pattern from the plurality of selectable hair patterns; 
 determining, based at least in part on the selection of the first hair pattern and from the first plurality of content items, a second plurality of content items that are associated with the first hair pattern; and 
 causing at least a portion of the second plurality of content items to be presented on the client device. 
 
     
     
       8. The computer-implemented method of  claim 7 , further comprising:
 determining, based at least in part on the query, that the query triggers hair pattern filtering of the first plurality of content items, and 
 wherein causing the hair pattern filter control to be presented on the client device is in response to the determination that the query triggers hair pattern filtering of the first plurality of content items. 
 
     
     
       9. The computer-implemented method of  claim 8 , wherein determining that the query triggers hair pattern filtering further includes, at least one of:
 determining that an inventory of content items of the second plurality of content items associated with at least one of the plurality of selectable hair patterns exceeds a threshold; 
 determining a relevance of the query to hair pattern filtering; or 
 determining that at least a portion of the query is included in a corpus of triggering queries. 
 
     
     
       10. The computer-implemented method of  claim 7 , further comprising:
 determining a diversification component associated with the first plurality of content items; and 
 causing at least a portion of the first plurality of content items to be presented on the client device in an arrangement based at least in part on the diversification component. 
 
     
     
       11. The computer-implemented method of  claim 7 , further comprising:
 determining a diversification component associated with the second plurality of content items; and 
 causing at least a portion of the second plurality of content items to be presented on the client device in an arrangement based at least in part on the diversification component. 
 
     
     
       12. The computer-implemented method of  claim 8 ,
 wherein each embedding vector of the plurality of embedding vectors is representative of an entirety of a corresponding content item of the corpus of content items. 
 
     
     
       13. The computer-implemented method of  claim 7 , wherein:
 at least one content item is associated with a plurality of hair patterns; and 
 the respective hair pattern associated with the at least one content item is a dominant hair pattern of the plurality of hair patterns. 
 
     
     
       14. The computer-implemented method of  claim 8 , wherein the determination of the corresponding hair patterns is performed without performing preprocessing on each content item of the corpus of content items. 
     
     
       15. A computer-implemented method, comprising:
 accessing a plurality of embedding vectors, wherein each embedding vector of the plurality of embedding vectors is representative of an entirety of a corresponding content item of a corpus of content items; 
 processing, using a trained machine learning model and without performing preprocessing of each corresponding content item, each embedding vector of the plurality of embedding vectors to determine a respective hairstyle for each corresponding content item; 
 associating each respective hairstyle with each corresponding content item; 
 determining, based at least in part on a query received from a client device, a first plurality of content items from the corpus of content items that are responsive to the query; 
 causing, in response to the query, at least a portion of the first plurality of content items and a hairstyle filter control that includes a plurality of selectable hairstyles to be presented on the client device; 
 obtaining, via an interaction with the hairstyle filter control, a selection of a first hairstyle from the plurality of selectable hairstyles; 
 determining a second plurality of content items from the first plurality of content items, wherein each of the second plurality of content items is associated with the first hairstyle; and 
 causing, in response to the selection of the first hairstyle, at least a portion of the second plurality of content items to be presented on the client device. 
 
     
     
       16. The computer-implemented method of  claim 15 , wherein the hairstyle includes at least one of a protective hairstyle, a coily hairstyle, a curly hairstyle, a wavy hairstyle, a straight hairstyle, or a shaved hairstyle. 
     
     
       17. The computer-implemented method of  claim 15 , wherein:
 determining, based at least in part on the query, that the query triggers hairstyle filtering of the first plurality of content items, and 
 causing the hairstyle filter control to be presented on the client device is in response to the determination that the query triggers hairstyle filtering of the first plurality of content items. 
 
     
     
       18. The computer-implemented method of  claim 17 , wherein determining that the query triggers hairstyle filtering further includes, at least one of:
 determining that an inventory of content items of the second plurality of content items associated with at least one of the plurality of selectable hairstyles exceeds a threshold; 
 determining a relevance of the query to hairstyle filtering; or 
 determining that at least a portion of the query is included in a corpus of triggering queries. 
 
     
     
       19. The computer-implemented method of  claim 15 , further comprising:
 obtaining, via a second interaction with the hairstyle filter control, a second selection of a second hairstyle from the plurality of selectable hairstyles; and 
 determining the second plurality of content items from the first plurality of content items is further based on the second hairstyle, such that each of the second plurality of content items is associated with at least one of the first hairstyle or the second hairstyle. 
 
     
     
       20. The computer-implemented method of  claim 15 , wherein at least one of:
 presentation of at least the portion of the first plurality of content items is based at least in part on a first diversity component associated with the first plurality of content items; or 
 presentation of at least the portion of the second plurality of content items is based at least in part on a second diversity component associated with the second plurality of content items.